[ATM] PlOp PlOp Fizz Fizz oh what a relief it isn't? Mirror cell for an 8" mirror

[rocketmaker@intergate.com]

Currently thanks to a short cable I am without my modeling software.  
Uhm I need a new screen for a Dell Inspiron 1000. Meanwhile I have  
some questions about the questions and answers given here regarding  
mirror cell design and the software PlOp. I hear all these definite  
statements and I would like to asks what is the validation for the  
software? I am familiar with Finite Element Modeling, I have also seen  
this technique applied by extremely large telescope manufacturers with  
much different results. So here are some thought questions and so forth.

I suspect PlOp is inadequate and produces false results that are  
inversely proportional to the number of supports. Also the old joke  
that big secondaries hide big central errors has been given a new face  
it seems. Let?s be accurate, secondaries do not affect the figure of  
the primary, it may sometimes mask its errors but that mainly applies  
to on axis testing and not the real world.  Then the concept of  
refocusing to obtain correction leads me to ask what happens when the  
telescope is pointing 60° from vertical. The idea should be to provide  
support so the mirror does not sag, not so it sags into an acceptable  
figure. Otherwise , why not just figure to a sphere, ring support and  
let the objective sag to correction?

I haven?t dissected the code but I did not see where the software  
dealt with boundaries. I did not see trans lateral vectors, there is a  
lot that seems missing at this point. In dealing with the eight inch  
mirror that is 15% thicker on the edge than the center. That means the  
mesh elements need to represent 15% more weight on the edge than the  
center.  I can produce a model that seems to agree with a 40% zone for  
3 point support but then I can produce one that is more complete that  
suggest 80%.  For D the 50% zone would be  ((D^2)/8)^½) or  D/2*2^½   
or approximately 71% of the diameter. Thus the idea would be that an  
annular ring at the 71% zone would support the optic. Of course a  
central perforation and thicker edge would shift the ring outwards.   
Three points may find level but they do not make a good support. On  
the other hand they may make an adequate support. My 6x¾? mirror has  
been in its cell for  38 years sitting on 3 points at 71% and still  
has an excellent figure.  One person said something about nine points  
not be good enough, really??? Has anyone ever seen a nine point cell  
fail? I have seen a few glued down mirrors warp but the main problem I  
have seen is from inadequate lateral edge support. I agree that some 9  
pt. cells are not designed as well as the could be but then there are  
mirrors that have been in nine point cells for  over 100yrs. without a  
problem.

So what validates PlOp? The web page shows simulated images from PlOp  
which we know looks nothing like the real thing so is there any  
empirical or anecdotal evidence that PlOp is right? I can guarantee  
you that if you designed a roof support using PlOp it would be turned  
down at the planning dept. A general rule for a cantilever requires  
two for one supported vs. overhang. In other words the support would  
be no further than 1/3 of the way in or at the 67% zone.

I could make a thin mirror that will show deflections more readily but  
then  actually there is an old mirror makers trick, That is to take a  
short fiber low or no grain (Kraft, or filter paper the same diameter  
as the mirror and drape it over the supports. This will give an  
indication of what the glass wants to do. You will see my point that  
PlOp allows the mirror to sag in a way that is tolerable by refocusing  
but will not hold focus or correction in attitudes from horizontal. I  
am always open minded but as I see it PlOp is not a real world  
modeling tool so has someone proof otherwise?

Telescope Maker,
Rick Crockett


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


_______________________________________________
ATM mailing list http://www.atmlist.net/